[Gimp-print-devel] Ghostscript driver

[Robert L K. <rl...@al...>]

It works!

Well, OK, it's still printing, and it will be for quite a while, but
it's definitely doing its thing here.  It's started printing the
second page (the dithering routine did the whole thing already), so
I'm quite confident that it's working.  There's no sign of any memory
leak; while the memory consumed was excessive (80 MB), I suspect that
that's Ghostscript's fault rather than the driver's.  During the
entire rendering phase the working set was about 4 MB, which is about
right by my back of the printhead calculations.

The bad news: performance is disastrous, both rendering and dithering.
It took over 16 minutes of CPU time on a reasonably fast machine (450
MHz Celeron, approximately equivalent to a 450 MHz PII) to render a 15
page document (admittedly, this is at 720 highest quality, which isn't
the fastest mode, but it's still faster than 1440).  I haven't hit any
color stuff yet, so we'll see how that does.

The driver definitely doesn't optimize out blank parts of the page,
which is important for good performance.

Anyway, I'm basically a happy camper right now.  We can always fix
performance.

(What's really ironic in a sense is that the document I'm printing out
is some what's new for Open Windows that looks like it was some slide
show generated by Sun back in 1993.  I work for Sun in a performance
group.  I should be embarrassed.)

This has me thinking the R-word, for obvious reasons.  What would
people like to see done before we do a development release?

-- 
Robert Krawitz <rl...@al...>      http://www.tiac.net/users/rlk/

Tall Clubs International  --  http://www.tall.org/ or 1-888-IM-TALL-2
Member of the League for Programming Freedom -- mail lp...@uu...
Project lead for The Gimp Print --  http://gimp-print.sourceforge.net

"Linux doesn't dictate how I work, I dictate how Linux works."
--Eric Crampton